Computer Vision and Pattern Recognition: Temporal Global Re-Detection Based on Interaction-Fusion Attention in Long-Term Visual Tracking -- VLNet: A Multi-task Network for Joint Vehicle and Lane Detection -- Adaptive Cost Aggregation in Iterative Depth Estimation for Efficient Multi-View Stereo -- A Novel Semantic Segmentation Method for High-Resolution Remote Sensing Images Based on Visual Attention Network -- Efficient Few-shot Image Generation via Lightweight Octave Generative Adversarial Networks -- Incorporating Global Correlation and Local Aggregation for Efficient Visual Localization -- Deep Interactive Image Semantic and Instance Segmentation -- Learning High-Performance Spiking Neural Networks With Multi-Compartment Spiking Neurons -- Attribute Space Analysis for Image Editing.-SAGAN: Self-Attention Generative Adversarial Net-work for RGB-D Saliency Prediction -- Behavioural State Detection Algorithm for Infants and Toddlers Incorporating Multi-scale Contextual Features -- Motion-Scenario Decoupling for Rat-Aware Video Position Prediction: Strategy and Benchmark -- Dual Fusion Network for Hyperspectral Semantic Segmentation -- Strip-FFT Transformer for single image deblurring -- Vision-Language Adaptive Mutual Decoder for OOV-STR -- DensityLayout: Density-conditioned Layout GAN for Visual-textual Presentation Designs -- GLTCM: Global-Local Temporal and Cross-Modal Network for Audio-Visual Event Localization -- Recent Advances in Class-Incremental Learning -- TTA-GCN: Temporal Topology Aggregation for Skeleton-Based Action Recognition -- A Stable Long-Term Tracking Method For Group-Housed Pigs -- Dynamic Attention for Isolated Sign Language Recognition with Reinforcement Learning -- A Segmentation Method based on SE Attention and U-Net for Apple Image -- Human Action Recognition Method based on spatio-temporal relationship -- Facial expression recognition from occluded images using deep convolution neural network with Vision Transformer -- Learning Discriminative Proposal Representation for Multi-Object Tracking -- Virtual-Violence: A Brand-New Dataset for Video Violence Recognition -- A novel Attention-DeblurGAN-Based Defogging Algorithm -- Multi-Modal Context-Aware Network for Scene Graph Generation -- VQA-CLPR: Turning a Visual Question Answering Model into a Chinese License Plate Recognizer -- Unsupervised Vehicle Re-Identification via Raw UAV Videos -- Distance-Aware Vector-Field and Vector Screening Strategy for 6D Object Pose Estimation -- SSTA-Net: Self-supervised Spatio-Temporal Attention Network for Action Recognition -- Gesture recognition method based on Sim-ConvNeXt model -- Research on Airborne Infrared Target Recognition Method based on Target-Environment Coupling -- Semantic and Gradient Guided Scene Text Image Super-Resolution.

The five-volume set LNCS 14355, 14356, 14357, 14358 and 14359 constitutes the refereed proceedings of the 12th International Conference on Image and Graphics, ICIG 2023, held in Nanjing, China, during September 22-24, 2023. The 166 papers presented in the proceedings set were carefully reviewed and selected from 409 submissions. They were organized in topical sections as follows: computer vision and pattern recognition; computer graphics and visualization; compression, transmission, retrieval; artificial intelligence; biological and medical image processing; color and multispectral processing; computational imaging; multi-view and stereoscopic processing; multimedia security; surveillance and remote sensing, and virtual reality. The ICIG 2023 is a biennial conference that focuses on innovative technologies of image, video and graphics processing and fostering innovation, entrepreneurship, and networking. It will feature world-classplenary speakers, exhibits, and high quality peer reviewed oral and poster presentations.